Earth and Planetary Science Letters 248 (2006) 451–461 www.elsevier.com/locate/epsl Cosmogenic and nucleogenic 3He in apatite, titanite, and zircon K.A. Farley a,⁎, J. Libarkin b, S. Mukhopadhyay c, W. Amidon a a Division of Geological and Planetary Sciences, California Institute of Technology, MS 170-25, Caltech, Pasadena, CA 91125, USA b Department of Geological Sciences, Ohio University, Athens, OH 45701, USA c Department of Earth and Planetary Sciences, Harvard University, Cambridge, MA 02138, USA Received 7 April 2006; received in revised form 6 June 2006; accepted 6 June 2006 Available online 17 July 2006 Editor: R.W. Carlson Abstract Cosmogenic 3He was measured in apatite, titanite, and zircon and cosmogenic 21Ne in quartz at 13 depth intervals in a 2.7-m long drill core in a Miocene ignimbrite from the Altiplano of Bolivia. All three 3He depth profiles as well as the 21Ne profile attenuate exponentially with depth, indicating that both of these isotopes are cosmogenic in origin with no significant contribution from other sources. The attenuation lengthscale for 3He production of Λ=180±11 g/cm2 is consistent with expectations for neutron spallation, and is identical to that found for the cosmogenic 21Ne in quartz. By normalizing the measured 3He concentrations to 21Ne and using the independently known cosmogenic 21Ne production rate, the apparent cosmogenic 3He production rates in apatite, titanite, and zircon were respectively found to be 112, 97, and 87 atoms/g/yr at sea-level and high latitude. The formal uncertainty on these estimates is ∼ 20% (2σ), and arises in equal parts from uncertainties in the measured 3He/21Ne ratios and the uncertainty in the 21Ne production rate. However an additional factor affecting the apparent 3He production rate in these phases arises from the long stopping range of spalled 3He and tritium (which decays to 3He). Because all three accessory phases have higher mean atomic number than major rock-forming minerals, they will have lower 3He production rates than their surroundings. As a consequence the long stopping ranges will cause a net implantation of 3He and therefore higher apparent production rates than would apply for purely in-situ production. Thus these apparent production rates apply only to the specific grain sizes analyzed. Analysis of sieved zircon aliquots suggests that a factor of 2 increase in grain size (from ∼ 50 to ∼ 100 μm cross-section) yields a 10% decrease in apparent production rate. While this effect warrants further study, the grain sizes analyzed here are typical of the accessory phases commonly encountered, so the apparent rates provide an appropriate starting place for surface exposure dating using 3He in these minerals. © 2006 Elsevier B.V. All rights reserved. Keywords: cosmogenic helium; apatite; zircon; titanite; production rate; Altiplano 1. Introduction 21Ne despite the comparatively good detection limits and the ease with which noble gases can be measured. In the The last few years have seen rapid development of case of 21Ne, the major limitation is nucleogenic neon from surface exposure dating using the cosmogenic radio- UandThdecay[4].For3He, there are two recognized nuclides 10Be, 26Al, and 36Cl [1–3].Lessworkhasbeen limitations. Non-cosmogenic helium is commonly found in done with stable cosmogenic isotopes such as 3He and minerals either from the reaction 6Li(n,α)3H→ 3He or from the presence of mantle helium in fluid inclusions [4]. More ⁎ Corresponding author. importantly, many rock-forming minerals do not retain E-mail address: [email protected] (K.A. Farley). helium against diffusion at earth surface temperatures. For 0012-821X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2006.06.008 452 K.A. Farley et al. / Earth and Planetary Science Letters 248 (2006) 451–461 example, quartz [5] and feldspar [6] leak helium at 20 °C, tion. The analyzed grains were generally small (∼ 50 μm while biotite and hornblende commonly have 10s of ppm in cross-section) and were a mixture of euhedral crystals of Li [7], which can yield substantial nucleogenic 3He and fragments. The titanite separate was sieved at 150 over time. These factors have restricted surface exposure μm, and the coarse fraction was handpicked to exclude dating using 3He almost exclusively to young basalts contaminants. The average cross-section of the analyzed bearing olivine or pyroxene phenocrysts (e.g., [4,8]). grains was 250 μm; most were fragments of originally More recently other mineral phases have come under larger grains. The zircon separate was heavily contam- investigation for 3He-based exposure dating, including inated with titanite, so it was placed in a mixture of 2:1 garnet [9] and magnetite [6]. Here we expand on this concentrated HF/HNO3 and heated to boiling in an open work by establishing a framework for surface exposure beaker for 1 h. The acid was replaced with concentrated dating using cosmogenic 3He in three accessory phases: HCl and heated for an additional hour. This heating will apatite (Ca5(PO4)3F), titanite (“sphene”, CaTiSiO5) and not release He from zircon [13]. The resulting separate zircon (ZrSiO4). These minerals are ubiquitous and re- was found to be pure zircon. Individual euhedral grains tain He under earth surface conditions [10]. Cosmogenic averaging ∼ 110 μm in cross-section were handpicked 3He has been detected in igneous apatites from the Dry and analyzed. After these samples were analyzed we Valleys of Antarctica, and in fossil tooth enamel apatite recognized a possible role of grain size on the cosmo- [11], but there appear to be no reports of cosmogenic genic 3He production rate, so for one sample (D1_D2) helium in titanite or zircon. Here we describe results we obtained a second zircon sample and sieved it into from a 2.7-m long depth profile of 3He measurements in two size fractions. Based on measurements of at least these three phases. Our goal is to establish that cosmo- 100 grains the mass-weighted mean grain cross section genic 3He is present in these minerals, to assess the of the coarse fraction was 91 μm, and for the fine frac- mechanism by which it is produced, and to obtain an tion, 53 μm. Most but not all grains were euhedral prisms estimate of the 3He production rates. We then consider rather than fragments. several factors which may complicate cosmogenic Helium was extracted by laser heating of 2 to 10 mg of dating of these phases. separate contained in 6-mm lengths of 3-mm OD Pt– 10%Ir tubing (Johnson Matthey Medical P/N 10170). 2. Samples and methods Prior to use the tubing was annealed and degassed by in- vacuum laser heating to incandescence. Each length of We analyzed samples from a core drilled into a thick tubing can hold about 2 mg of separate, so larger aliquots ignimbrite on the Bolivian Altiplano. The ignimbrite was were divided among multiple tubes. After loading, the selected for its high abundance of apatite, titanite, and tube ends were crimped and the aliquots were placed into zircon, plus quartz for 21Ne measurement. The ignimbrite a vacuum chamber below a sapphire viewport. Follow- was sampled at 18° 15.694′ S, 68° 14.702′ W, elevation ing 12 h of evacuation each aliquot was heated with a 3843 m. It likely corresponds to sample 90BR037 reported Nd-YAG laser to ∼ 1300 °C for 15 min. During this by du Bray et al. [12], who provide major element data for period the laser power was controlled by monitoring the this ignimbrite. As discussed below, it is Miocene in age brightness of the packet to maintain the desired temper- and has an apparent exposure age based on 21Ne in quartz ature, and the beam was rastered across the tubes to of ∼ 100 kyr, indicating significant erosion. The ignim- ensure uniform degassing. All helium is extracted from brite is a moderately welded rhyolite ash-flow forming a these minerals at this temperature [10] as confirmed by a laterally extensive, flat, continuously exposed surface. A second heating and analysis of several samples. The total of 2.7 m of core was drilled in sections ranging in evolved helium was purified, cryofocussed, and sepa- length from 2 to 40 cm, with drilling-related loss of ap- rated from Ne using standard procedures at Caltech [14]. proximately 40 cm from the lower two-thirds of the core. The helium was admitted into a MAP 215-50 mass Depths above 150 cm are exact to ±5 cm, and depths spectrometer and the two isotopes analyzed by peak- below are exact to ±10 cm. Core sections were cut into jumping. 3He was detected on a pulse-counting electron 2–4 cm pieces, from which 13 samples were processed multiplier. for heavy minerals. Laser extraction blanks were close to the detection Segments of core totaling 2 to 10 cm in length were limit for 3He (∼ 3000 atoms) and were <2×109 atoms crushed to pea-gravel size, and apatite, titanite and zircon for 4He. In all cases the 4He blanks were negligible were purified using standard heavy liquid and magnetic (<1%). The 3He blank was ≪1% for most samples, techniques. The resulting apatite separate was visually reaching a maximum of 5% on the samples from deepest pure and aliquots were analyzed without further prepara- in the core. The 3He abundance was determined by peak K.A. Farley et al. / Earth and Planetary Science Letters 248 (2006) 451–461 453 height comparison with gas standards. Linearity of 3He correction. Cosmogenic 21Ne concentrations were cal- peak heights was established by analyzing standards of culated assuming a two-component mixture of cosmo- known size; these standards encompassed the amount of genic and air neon.